Page History

Product info here (rack-mounted version here)

Brain dump:

• We have two of these, with the rack-mounted version serving as back-up; located in rack 64B; the back-up has been tested for functionality previously, so it should be plug and play
• These can be wavelength tuned, but it's a little finnicky and laggy (especially on the rack-mounted version, which is tuned using a pot on the front of the unit), so it's best not to touch them; in the ancient past, these units got shuffled around with the factory supposedly in order to get the tuning range set appropriately for the system it would be seeding
• There is a way to read out information from these units from the back panel, but there is no way to actually control these units from the back panel – no hope for writing an IOC
• Only plugs in the back should be power and interlock
• System should show temp stable and second mode suppressed or else there will be major performance/stability issues; depending on how long it's been off or how unstable the environment is, this may be fine right away or it may take ~30-60min
• Front display should show ~135-140mW and 1053.045nm when running
• Past experimenting didn't show any effect of using ACC vs. APC mode nor RIN suppression on vs. off
• Upgrade ideas:
  • This is essentially the only laser of its kind I know of that doesn't use a fiber pre-amplifier somewhere before the free-space amplifier. Sometimes this is located before or after the EOM (or both), but it can provide way better performance this way.

Representative/historical pictures:

Image AddedImage Added

Product info here

Brain dump:

• EOM is mounted on the mezzanine inside the Highland T400B AWG, with fibers coupled to the back panel.
• Besides the two fiber connections, the EOM has three electrical connections too:
  • Shaped RF input taken from the output of the T400B's AWG output
  • Bias voltage input taken from the output of the MBC
  • On-board photodiode signal out, which is usually not plugged in but its BNC connector is hanging out loose on the front side of the T400B; not very useful except when doing basic troubleshooting to see if there's any energy coming through the EOM at all
• The input power is rated to be <100mW, but they've said there shouldn't be any problem putting in what we've been doing. They aren't a particular useful lot when it comes to getting troubleshooting advice from them, but often it centers on trying to attenuate a bunch to see what happens – we've just found work-arounds for our problems instead, but maybe that could be done differently in the future with more time and effort.
• Once upon a time, Michael Greenberg got a TEC for temperature stabilizing the unit to help it run for stably, but especially after relocating the T400B into the controls rack the performance doesn't seem to be affected too much so we never put it together, even though Marcio or someone of the like wrote an IOC for the TEC at some point
• The extinction is supposed to provide at least a polarization contrast of >1000:1; however, we have tried measuring this in the past as we've noticed that there's more light than normal sneaking through the EOM when it's "off" and causing a pre-pulse/pedestal between the slow opening of the Pockels cell in the YFE and the true rising edge of the laser pulse. We don't have a good way to measure the amount of leakage at very low power levels, so I would calibrate the strength of signal at full throughput (MUCH lower than the ~140mW supposedly getting injected into the EOM) with a power meter, a CCD, and a calibrated filter attenuator. From there, I could estimate that at the zero output case the attenuation was closer to 23dB than >30dB.
• We used this measurement also to estimate the amount of pulse energy actually making it out of the EOM – it's probably on the 100pJ-scale.
• Upgrade ideas:
  • Many other facilities use two-stage EOMs instead of single-stage. In this case, there's a second input that can be used with a high-bandwidth square pulse (for example) to sharpen up the rising and falling edges of the optical pulse AND to improve the polarization extinction on the front edge. In fancier cases they can pull tricks to make a pulse that already has a bit of a rising edge in order to not waste pixel depth of the AWG on the front side of the pulse (i.e. if you know the rising edge needs to be at 10%, if your pulse starts at 12% then you have a lot more pixel range to tune with than if you start at 100% and need to get down to 10%).

Representative/historical pictures:

Image Added

Technical manual here

Brain dump:

• AWG is mounted in the same rack as the rest of the shaping hardware, with the EOM mounted inside on an internal mezzanine.
• Originally purchased as the same realization as what is used at LLNL with the NIF front-end systems.
• All of the control of the Highland is done through Python sockets directly over the network and a Python class written by Eric Cunningham. 
• In order to communicate with the Highland, your computer has to be on the same subnet as the Highland. The Highland's IP address can be set using a Lantronix XPort module whose software should be found here.
• Most of the use of this system is reading and writing pulse heights, but there are many other utilities possible such as control of a ~100ps fiducial pulse (usually used for co-timing of different legs compared to a window on an oscilloscope) and many internal settings. The internal settings shouldn't be messed with too much at all, but we have used these occasionally in the past to make sure that the impulses are equally separated or that the gain is relatively flat across the timing window. Hopefully this isn't necessary too often (there are some old rudimentary routines that exist in some files for these kinds of things that were never fully developed since it ended up not being needed frequently enough to be necessary), especially if the unit is kept in the same place and stays in a stable environment, but Highland has been responsive in the past to any questions that may pop up.
• Further questions should generally be answerable looking at the manual or at the Python code.
• NOTE: there may be a few errors in the manual, such as the READ/WRITE MISCELLANEOUS CALIBRATIONS reply/poll length being 81 rather than 79; proceed with caution; the Python code should work though – the learning should be baked in there
• Upgrade ideas:
  
  • The T400B is very old technology. There are much newer boxes that could provide higher pixel density, longer arbitrary waveform temporal shaping window, or even both (though there tends to be a trade-off between the two).
  • There is an IOC that was written for the Highland, but we don't use it because it proved to be too slow to operate at 10Hz for our updating loops, so we have it disabled right now. If this could be fixed, we wouldn't have to control it ourselves anymore.
  • There are unused capabilities with this AWG, though it might not be appropriate if the 

Product info here
User manual here
Interface manual here
Software here

Brain dump:

• Most of the bias control parameters can be found in the additional info above.
• The MBC provides the bias needed to operate the EOM at the zero-point such that any applied voltage from the AWG translates to intensity-modulated output.
• In order to detect where the appropriate operating point should be, the MBC applies a dither to the voltage output and monitors the drifts in the necessary voltage set-point. This is a MASSIVE problem because the dither clock is NOT locked to the house clock. This means that when the actual pulses are triggered, the EOM is likely NOT operating at the necessary voltage set-point because it is being constantly driven AWAY from that setpoint by the dither. The dither frequency and voltage can be controlled directly by the manufacturer's software, but it has its problems (the available settings range for voltage and frequency is limited, and we had issues with the software being able to run at all unless the OS's default numbering protocol was set to European style with a comma representing a decimal rather than a point), and these settings aren't accessible from the IOC.
  • As a result, shaping performance is very unstable and unpredictable UNLESS the bias dither is turned OFF for any high-consequence activities such as recipe creation/shaping and on-target (user) shots. For the short term, disabling the bias dither should not cause any degradation in performance – the drift being monitored by the dither generally takes place on the time scales of tens of minutes or hours. The dither can be turned off by switching from AUTO to MAN(ual) voltage control in the IOC.
  • HOWEVER, re-engaging the dither causes the voltage output to be scanned across the full output range of -10V to +10V. If CW light is seeded into the amplifiers, then high-energy microsecond pulses can be output that can damage the serrated aperture at the output of the YFE or other optics. As such, the currents of the YFE heads need to be turned down before the MBC is switched back from MAN to AUTO. The shaping code should do this automatically if controlling the MBC from there; if controlling it from the IOC directly, exercise caution!
• Upgrade ideas:
  
  • Solve the dither problem – either get it locked to the house clock or get rid of it completely!!

Product info here